One of the often observed sea-coastal area interactions is the land-sea breeze. That phenomenon is the result of different physical conditions of water and land surfaces in coastal regions. This study intends to discuss an uncertainty analysis performed on hydrological processes. The main interest is on the water cycle relevant quantities. The atmospheric water cycle strongly depends on heat and moisture fluxes (i.e. it is driven by soil-biosphere-atmosphere interactions). Since the area of interest is a coastal region, differences in soil and sea conditions are considered. Prediction of the sensible and latent heat fluxes over the land mass is quite uncertain due to necessity of prescribed plant physiological (leaf area index, stomatal resistance etc) and soil physical parameters (soil heat capacity, soil thermal and hydraulic conductivity etc.). Knowledge on uncertainty and sensitivity of parameters may improve the ability to predict the water cycle relevant quantities. The basic idea of this study is to estimate the errors in the predicted latent and sensible heat fluxes by Gaussian Error Propagation. Method considers linear variation of quantities predicted in the land surface scheme with respect to the prescribed parameters in order to find the range where land surface model simulations are the most accurate. It is expected that the effects of the uncertainty in the parameters on predicted quantities could vary substantially. This means that parameters affect the accuracy of simulations for some quantities differently at different soil, sea and atmospheric conditions. As opposed to the land surface, there is no such an uncertainty on these parameters over the sea surface. Differences in physical characteristics, fluxes of latent and sensible heat, between water and land masses lead to specific small-scale circulation which is known as the land sea breeze. Since heat and moisture fluxes over the water mass can be predicted with high certainty, it seems that accuracy of prediction of the whole process is controlled exclusively by forecasted quantities over the land mass. To get bigger picture and better knowledge of combinations of atmospheric, soil and sea conditions, the analysis is performed within the NCAR MM5 model.